An Efficient and Robust Deep Learning Approach to Predict Air Pollution by Employing Long Short-Term Memory

Predicting air pollutant concentrations is an important tactic for safeguarding human health and providing advanced warning of dangerous airborne pollutants. Unfortunately, conventional techniques of predicting air pollution concentrations fail to predict long-term dependencies properly and ignore geographical connections. The characteristics of air quality are typically mirrored by a variety of elements, including temperature, wind direction, humidity, wind speed, rainfall, snowfall, and so on, which complicates comprehending the shift in air pollutant intensity. In this study, a convolutional neural learning-based Long Short- Term Memory (LSTM) predicting algorithm for PM2.5 (air pollutants having an aerodynamic size less than or equal to
$2.5\mu$
m) intensity is proposed. Hourly datasets spanning from 2010 to 2014 are analysed to evaluate the overall efficacy of the suggested LSTM model. This research assesses the LSTM algorithm's practical value for a pollutants prediction system for health mitigating risk through performance standards and comprehensive defect detection.